CN110347810B - Dialogue type search answering method, device, computer equipment and storage medium - Google Patents
Dialogue type search answering method, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN110347810B CN110347810B CN201910464157.2A CN201910464157A CN110347810B CN 110347810 B CN110347810 B CN 110347810B CN 201910464157 A CN201910464157 A CN 201910464157A CN 110347810 B CN110347810 B CN 110347810B
- Authority
- CN
- China
- Prior art keywords
- knowledge
- keywords
- financial
- node
- keyword
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
- G06F40/295—Named entity recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a knowledge graph-based conversational retrieval answering method, a device, computer equipment and a storage medium, wherein the method comprises the steps of obtaining financial question sentences; semantically identifying a financial problem statement, and splitting the financial problem statement into a plurality of keywords; confirming node search paths of knowledge nodes in a financial knowledge graph according to the front and back sequence of a plurality of keywords in a financial problem statement; finding a corresponding answer knowledge node in the financial knowledge map according to the node search path, calling the answer content of the answer knowledge node, and feeding the answer content back to the computer equipment; therefore, the problem that the existing machine has no function of eliminating ambiguity and the technical problem of low unformed answer speed of the financial knowledge map is solved.
Description
Technical Field
The invention relates to the field of financial knowledge maps, in particular to a dialogue type retrieval and answer method, a dialogue type retrieval and answer device, computer equipment and a storage medium.
Background
At present, the financial knowledge map on the market is not formed; when a user searches for financial questions, a computer generally needs to search in the internet according to the questions spoken by the user, and answers the searched answers to the user, so that a financial knowledge map is not used, and the machine is slow in answering speed.
Meanwhile, when a user speaks an ambiguous utterance at the time of a conversation, the machine does not have a disambiguation function, resulting in that the machine cannot give an answer intended by the user.
Disclosure of Invention
The invention aims to provide a dialogue type retrieval answering method, a dialogue type retrieval answering device, computer equipment and a storage medium, and aims to solve the technical problems that a machine does not have a function of disambiguation and the unfinished financial knowledge map is slow in answering speed.
In order to achieve the purpose, the invention provides the following technical scheme:
the invention provides a dialogue type search answering method, which comprises the following steps:
acquiring a financial problem statement input by a user through computer equipment;
semantically identifying the financial problem statement, and splitting the financial problem statement into a plurality of keywords;
analyzing whether ambiguous keywords exist in the plurality of keywords, wherein the ambiguous keywords are keywords which can be found out from a plurality of knowledge nodes in a pre-established financial knowledge graph;
if yes, feeding back a plurality of knowledge nodes corresponding to the ambiguous keyword to the computer equipment so that the user can select one knowledge node as the knowledge node corresponding to the ambiguous keyword;
analyzing whether the plurality of keywords have invalid keywords which do not have corresponding knowledge nodes in the financial knowledge graph;
if yes, a new knowledge node corresponding to the invalid keyword is created in the financial knowledge graph, corresponding internet search is carried out according to the invalid keyword, and the searched keyword content is used as the node content of the new knowledge node;
determining knowledge nodes corresponding to normal keywords in the plurality of keywords in the financial knowledge graph; confirming corresponding node search paths according to the front-back sequence of the ambiguous keywords, the invalid keywords and the normal keywords in the financial problem sentence and the knowledge nodes respectively corresponding to the ambiguous keywords, the invalid keywords and the normal keywords; the normal keywords are keywords which can find the only knowledge node in the pre-established financial knowledge graph;
and finding a corresponding answer knowledge node in the financial knowledge map according to the node search path, calling the answer content of the answer knowledge node, and feeding the answer content back to the computer equipment.
Further, the financial knowledge map establishing method comprises the following steps:
acquiring text data input by a manager through a management terminal, wherein the text data comprises entity data and event data, the entity data is all knowledge nodes, and the event data is associated with different knowledge nodes;
performing association connection on the knowledge nodes through the event data to obtain a meshed primary knowledge graph;
searching the internet for all knowledge nodes in the primary knowledge graph to obtain knowledge node contents corresponding to all the knowledge nodes one by one;
and associating all knowledge nodes with the corresponding knowledge node content to obtain the financial knowledge graph.
Further, the step of acquiring the text data input by the manager through the management terminal includes:
identifying whether the same knowledge nodes with the same semantics exist in all the knowledge nodes;
and if so, integrating two or more same knowledge nodes into one knowledge node.
Further, the step of analyzing whether the plurality of keywords have invalid keywords in the financial knowledge graph without corresponding knowledge nodes comprises:
if yes, performing Internet search on the invalid keyword to judge whether the invalid keyword is an abbreviated keyword;
if yes, acquiring a complete keyword corresponding to the abbreviated keyword;
searching whether a knowledge node corresponding to the complete keyword exists in the financial knowledge graph;
and if the complete key words do not exist, taking the complete key words as the invalid key words, generating instructions for creating new knowledge nodes corresponding to the invalid key words in the financial knowledge graph, and performing corresponding internet search according to the invalid key words.
Further, the semantically recognizing the financial question sentence, and the step of splitting the financial question sentence into a plurality of keywords comprises:
splitting the financial problem statement into a plurality of words, and identifying keywords corresponding to the words respectively;
if the recognition is not available, searching whether a professional term corresponding to the unrecognized term exists in a professional term library, wherein the professional term library is a database in which all the professional terms are prestored;
and calling the professional term as a keyword corresponding to the unrecognized term, and taking the keyword corresponding to each of the plurality of terms as a plurality of keywords obtained by splitting the financial problem statement.
The invention provides a dialogue type search answering device, comprising:
the statement acquisition unit is used for acquiring financial problem statements input by a user through computer equipment;
the keyword identification unit is used for semantically identifying the financial problem statement and splitting the financial problem statement into a plurality of keywords;
the ambiguity analysis unit is used for analyzing whether an ambiguity keyword exists in the plurality of keywords, wherein the ambiguity keyword is a keyword which can search a plurality of knowledge nodes in a pre-established financial knowledge graph;
an ambiguity determining unit, configured to feed back, if yes, the knowledge nodes corresponding to the ambiguous keyword to the computer device, so that the user can select one of the knowledge nodes as the knowledge node corresponding to the ambiguous keyword;
an invalid analysis unit, configured to analyze whether an invalid keyword in the plurality of keywords exists in the financial knowledge graph, where the invalid keyword does not exist in a corresponding knowledge node;
an invalid determining unit, configured to create a new knowledge node corresponding to the invalid keyword in the financial knowledge graph if the invalid keyword exists, perform corresponding internet search according to the invalid keyword, and use the searched keyword content as the node content of the new knowledge node;
the path determining unit is used for determining knowledge nodes corresponding to normal keywords in the plurality of keywords in the financial knowledge graph; confirming corresponding node search paths according to the front-back sequence of the ambiguous keywords, the invalid keywords and the normal keywords in the financial problem sentence and the knowledge nodes respectively corresponding to the ambiguous keywords, the invalid keywords and the normal keywords; the normal keywords are keywords which can find the only knowledge node in the pre-established financial knowledge graph;
and the answer feedback unit is used for finding the corresponding answer knowledge node in the financial knowledge map according to the node search path, calling the answer content of the answer knowledge node and feeding the answer content back to the computer equipment.
Further, the dialogue-based search answering device further comprises:
the system comprises a data acquisition unit, a data processing unit and a data processing unit, wherein the data acquisition unit is used for acquiring text data input by a manager through a management terminal, the text data comprises entity data and event data, the entity data are all knowledge nodes, and the event data are associated with different knowledge nodes;
the preliminary forming unit is used for performing association connection on the knowledge nodes through the event data to obtain a meshed preliminary knowledge graph;
the content searching unit is used for searching the internet for all knowledge nodes in the primary knowledge graph in the mesh shape to obtain knowledge node contents which are in one-to-one correspondence with all the knowledge nodes;
and the content association unit is used for associating all the knowledge nodes with the corresponding knowledge node content to obtain the financial knowledge graph.
Further, the dialogue-based search answering device further comprises:
the same semantic identifying unit is used for identifying whether the same knowledge nodes with the same semantics exist in all the knowledge nodes;
and the integration unit is used for integrating two or more same knowledge nodes into one knowledge node if the knowledge nodes are the same.
The invention also provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned dialogue-based search and answer method when executing the computer program.
The present invention also provides a computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the above-mentioned conversational search answering method.
The invention provides a dialogue type retrieval answering method, a dialogue type retrieval answering device, a computer device and a storage medium, which have the following beneficial effects:
the server acquires a financial question sentence input by a user through computer equipment; semantically identifying a financial problem statement, and splitting the financial problem statement into a plurality of keywords; analyzing whether ambiguous keywords exist in the plurality of keywords, wherein the ambiguous keywords are keywords of which a plurality of knowledge nodes can be found in a pre-established financial knowledge map; if yes, feeding back a plurality of knowledge nodes corresponding to the ambiguous key words to the computer equipment so that the user can select one knowledge node as the knowledge node corresponding to the ambiguous key words; analyzing whether the plurality of keywords have invalid keywords which do not have corresponding knowledge nodes in the financial knowledge graph; if so, creating a new knowledge node corresponding to the invalid keyword in the financial knowledge graph, performing corresponding internet search according to the invalid keyword, and taking the searched keyword content as the node content of the new knowledge node; determining knowledge nodes corresponding to normal keywords in a plurality of keywords in a financial knowledge graph; confirming corresponding node search paths according to the front and back sequence of the plurality of keywords in the financial problem sentence and the knowledge nodes corresponding to the keywords respectively; and finding the corresponding answer knowledge node in the financial knowledge map according to the node search path, calling the answer content of the answer knowledge node, and feeding the answer content back to the computer equipment, so that the problem that the machine has no function of eliminating ambiguity and the technical problem that the financial knowledge map is not formed and has low answer speed is solved.
Drawings
FIG. 1 is a flow chart illustrating a method for dialogue-based search and answer according to a first embodiment of the present invention;
FIG. 2 is a flowchart illustrating a conversational search answering method according to a second embodiment of the invention;
FIG. 3 is a block diagram schematically showing the structure of a dialogue-based search answering apparatus according to a first embodiment of the present invention;
fig. 4 is a block diagram schematically showing the construction of a dialogue-based search response apparatus according to a second embodiment of the present invention;
FIG. 5 is a block diagram illustrating the structure of an embodiment of the computer apparatus of the present invention.
The implementation, functional features and advantages of the present invention will be further described with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention provides a server, which executes a dialogue type search answering method.
Referring to fig. 1, a flow chart of a conversational search answering method provided by the invention includes:
s100, acquiring a financial question sentence input by a user through computer equipment;
the financial question sentence is a text sentence input by a user through operating a computer device, and comprises a text sentence formed by recognizing voice information of the user and a text sentence manually input by the user. Financial problems in particular, such as: and B, tax data of C year in city B of province A, if the financial question sentences are input by the user through computer equipment voice, the server performs natural language processing on the financial question voice to obtain text-form financial question sentences.
S200, semantically identifying a financial problem sentence, and splitting the financial problem sentence into a plurality of keywords;
to remove useless text in the financial question sentence, for example: the financial question statement is "tax data and GPD growth rate of year C, B, a province? The steps of obtaining the keywords of ' province A, city B, year C, tax data and GDP growth rate ' after removing useless text ' are three:
the first is to identify the semantics of the financial question sentence and only keep the nouns in the financial question sentence, so as to obtain a plurality of keywords.
Secondly, pre-storing preset keywords preset by a manager in a keyword storage library of the server, and if the same keywords as the preset keywords appear in the financial question sentences, reserving the keywords so as to obtain a plurality of keywords;
and thirdly, finding out key words in the financial problem sentences through the knowledge nodes in the financial knowledge graph, wherein the knowledge nodes have knowledge nodes of province A, city B, year C, tax data and GDP growth rate, and thus obtaining a plurality of key words.
S300, analyzing whether ambiguous keywords exist in the plurality of keywords, wherein the ambiguous keywords are keywords which can be found out from a plurality of knowledge nodes in a pre-established financial knowledge map;
the financial knowledge graph is a knowledge graph which is created in advance and stored in the server, and the server judges whether ambiguous keywords of a plurality of keywords corresponding to a plurality of knowledge nodes exist in the financial knowledge graph or not after acquiring the keywords. For example: x is common, knowledge nodes of common budget, common budget expenditure and common budget income are stored in the financial knowledge map, so that keyword X in the financial problem statement has three knowledge nodes in the financial knowledge map, and the keyword X is an ambiguous keyword.
S400, if yes, feeding back a plurality of knowledge nodes corresponding to the ambiguous keywords to computer equipment so that a user can select one knowledge node as a knowledge node corresponding to the ambiguous keywords;
if the server judges that one keyword corresponds to a plurality of knowledge nodes, the server calls the knowledge nodes in the form of options and sends the options corresponding to the knowledge nodes to the computer equipment in the form of texts, so that the options can be conveniently presented on a screen of the computer equipment, and the effect of eliminating keyword ambiguity can be achieved by selecting one knowledge node by a user. For example: and if the keyword X is common, the server calls a common budget, a common budget expenditure and a common budget income corresponding to the keyword X from the financial knowledge map, and feeds the common public budget, the common public budget expenditure and the common public budget income back to the computer equipment of the user to select any item, so that the effect of eliminating the ambiguity of the keyword X is achieved.
S500, analyzing whether the plurality of keywords have invalid keywords of corresponding knowledge nodes which do not exist in the financial knowledge graph;
the server analyzes whether the keywords have invalid keywords which do not exist in the financial knowledge graph, and the mode is as follows: and searching whether knowledge nodes with the same meaning as the keywords exist in the financial knowledge graph. For example: the keyword "guangzhou city, tax payment data" appears in the financial problem statement, but only "guangzhou city" exists in the financial knowledge map but not "tax payment data", so the keyword "tax payment data" is an invalid keyword.
S600, if so, creating a new knowledge node corresponding to the invalid keyword in the financial knowledge graph, performing corresponding Internet search according to the invalid keyword, and taking the searched keyword content as the node content of the new knowledge node;
if the server recognizes that the plurality of keywords have invalid keywords, the invalid keywords are added to the financial knowledge graph as new knowledge nodes, but the invalid keywords are not associated with any content, so that the server searches for the content corresponding to the invalid keywords through the Internet to obtain node content, and the node content is used as the content of the invalid keywords, so that the financial knowledge graph is expanded and trained in such a way.
For example: and taking the keyword 'tax payment data' as an invalid keyword, adding a new knowledge node of the 'tax payment data' in the financial knowledge map, searching the content corresponding to the 'tax payment data' in the Internet, taking the searched content as the node content, and associating the node content with the knowledge node of the 'tax payment data'.
S700, determining knowledge nodes corresponding to normal keywords in a plurality of keywords in a financial knowledge graph; confirming corresponding node search paths according to the front-back sequence of the ambiguous keywords, the invalid keywords and the normal keywords in the financial problem sentences and the knowledge nodes respectively corresponding to the ambiguous keywords, the invalid keywords and the normal keywords; the normal keywords are keywords which can find the unique knowledge nodes in a pre-established financial knowledge graph;
as shown in the above example, the sequence of the keywords in the financial problem statement is "A province, B city, C year, and tax data", so the server finds the node search path of the knowledge node corresponding to the "A province, B city, C year, and tax data" in the financial knowledge graph.
And S800, finding the corresponding answer knowledge node in the financial knowledge map according to the node search path, calling the answer content of the answer knowledge node, and feeding the answer content back to the computer equipment.
As shown in the above example, the node search path of the keywords in the fiscal question sentence is "tax data" in a, B, C, so that the answer knowledge node is "tax data", and the server retrieves the answer content of "tax data" and feeds the answer content back to the computer device, wherein the answer content includes one or more of text answers, table answers, or picture answers.
Acquiring a financial question sentence input by a user operating computer equipment; semantically identifying a financial problem statement, and splitting the financial problem statement into a plurality of keywords; analyzing whether an ambiguous keyword of a keyword corresponding to a plurality of knowledge nodes exists in a pre-established financial knowledge graph in the plurality of keywords; if yes, feeding the knowledge nodes back to the computer equipment so that the user can select one knowledge node as a knowledge node corresponding to the ambiguous key word; analyzing whether the plurality of keywords have invalid keywords which do not exist in the financial knowledge graph; if so, creating a new knowledge node corresponding to the invalid keyword in the financial knowledge graph, performing corresponding internet search according to the invalid keyword, and taking the searched keyword content as the node content of the new knowledge node; confirming node search paths of knowledge nodes in a financial knowledge graph according to the front and back sequence of a plurality of keywords in a financial problem statement; finding a corresponding answer knowledge node in the financial knowledge map according to the node search path, calling the answer content of the answer knowledge node, and feeding the answer content back to the computer equipment; therefore, the problem that the existing machine has no function of eliminating ambiguity and the technical problem that the financial knowledge map is not formed and has low answer speed is solved.
Referring to fig. 2, a flow chart of the interactive search answering method according to an embodiment of the present invention is schematically illustrated, and a method for creating a financial knowledge base map is provided, including:
s001, acquiring text data input by a manager through a management terminal, wherein the text data comprises entity data and event data, the entity data is all knowledge nodes, and the event data is knowledge nodes associated with different associations;
the manager is a user who makes a financial knowledge map through the server and manages the server; the entity class data comprises a first-order entity, a second-order entity, a third-order entity and the like, the event class data is a connection bridge of the entity class data, the first-order entity is connected with the second-order entity through an event, and the first-order entity is connected with the other first-order entity through an event.
For example: the first-order entity is 'A province', the second-order entity is 'B city', the event class data comprises an event type which is 'lower city', and therefore the events of 'A province' passing through 'lower city' are related to 'B city' in the financial knowledge map. Need to explain: the entity class data are knowledge nodes, and the event class data are conditions for associating the knowledge nodes.
S002, performing association connection on the knowledge nodes through the event data to obtain a meshed primary knowledge graph;
s003, searching the Internet aiming at all knowledge nodes in the primary knowledge graph of the mesh to obtain knowledge node contents corresponding to all the knowledge nodes one by one;
and S004, associating all knowledge nodes with corresponding knowledge node contents to obtain a financial knowledge map.
After the server makes the preliminary knowledge graph, all knowledge nodes need to be filled with contents, and the method is as follows: searching all knowledge nodes one by one through the Internet to obtain the contents of the knowledge nodes corresponding one by one; and finally, performing association binding on the knowledge node content and the knowledge node corresponding to the knowledge node content to obtain the financial knowledge graph.
In another embodiment, the financial knowledge-graph is regenerated within a predetermined time period in order to update the data, such as once a month.
In one embodiment, the step S001 of acquiring the text data input by the manager through the management terminal includes:
s011 identifies whether the same knowledge nodes with the same semantics exist in all the knowledge nodes;
and if S021 is true, integrating two or more same knowledge nodes into one knowledge node.
In the financial knowledge graph making process, if the server identifies that the same knowledge node with the same semantics exists in all knowledge nodes, for example: and obtaining two knowledge nodes of the government of province A and the civil government of province A, identifying the semantics of the two knowledge nodes by the server, and integrating the two knowledge nodes into one knowledge node if the semantics are the same, namely the government of province A.
In one embodiment, the step S400 of analyzing whether the keywords have invalid keywords corresponding to the knowledge nodes in the financial knowledge graph includes:
s410, if yes, carrying out Internet search on the invalid keywords to judge whether the keywords are abbreviated keywords;
s420, if yes, acquiring a complete keyword corresponding to the abbreviation keyword;
s430, searching whether a knowledge node corresponding to the complete keyword exists in the financial knowledge graph;
and S440, if the key words do not exist, taking the complete key words as invalid key words, generating an instruction for creating new knowledge nodes corresponding to the invalid key words in the financial knowledge graph, and performing corresponding Internet search according to the invalid key words.
For example: the financial problem statement input by the user is 'high and new enterprises in Shenzhen city', wherein the 'high and new enterprises' do not have corresponding knowledge nodes in the financial knowledge map, so that the 'high and new enterprises' are subjected to internet search, whether the 'high and new enterprises' are abbreviated keywords is judged, if the server searches for corresponding complete keywords through internet search, the complete keywords are obtained, and if the 'high and new enterprises' are subjected to internet search, the 'high and new technical enterprises' are obtained. And the server searches whether the knowledge node corresponding to the complete keyword exists in the financial knowledge graph or not, if not, the complete keyword is used as an invalid keyword to generate a new knowledge node for creating the corresponding invalid keyword in the financial knowledge graph, and a creating instruction for searching the corresponding Internet is carried out according to the invalid keyword.
In one embodiment, the step S200 of semantically recognizing the financial question sentence and splitting the financial question sentence into a plurality of keywords comprises:
s210, splitting the financial problem sentence into a plurality of words, and identifying keywords corresponding to the words respectively;
s220, if the recognition is not available, searching whether a professional term corresponding to the unrecognized term exists in a professional term library, wherein the professional term library is a database in which all the professional terms are prestored;
and S230, calling the professional term as a keyword corresponding to the unrecognized term, and taking the keyword corresponding to each term as a plurality of keywords obtained by splitting the financial problem statement.
The above-mentioned unidentifiable keywords are keywords that the server cannot obtain through semantic processing, so it is necessary to search the corresponding professional term in the professional term library.
Referring to fig. 3, a schematic block diagram of a dialogue-based search answering device provided by the present invention includes:
a sentence acquisition unit 10 for acquiring a financial question sentence input by a user through a computer device;
the financial question sentence is a text sentence input by a user through operating a computer device, and comprises a text sentence formed by recognizing voice information of the user and a text sentence manually input by the user. Financial problems in particular, such as: and B, tax data of C year in city B of province A, if the financial question sentences are input by the user through computer equipment voice, the server performs natural language processing on the financial question voice to obtain text-form financial question sentences.
A keyword recognition unit 20, configured to semantically recognize a financial question sentence, and split the financial question sentence into a plurality of keywords;
to remove useless text in the financial question sentence, for example: the financial question statement is "tax data and GPD growth rate of year C, B, a province? The steps of obtaining the keywords of ' province A, city B, year C, tax data and GDP growth rate ' after removing useless text ' are three:
the first is to recognize the semantics of the financial question sentence and only keep the nouns in the financial question sentence, so as to obtain a plurality of keywords.
Secondly, pre-storing preset keywords preset by a manager in a keyword storage library of the server, and if the same keywords as the preset keywords appear in the financial question sentences, reserving the keywords so as to obtain a plurality of keywords;
and thirdly, finding out key words in the financial problem sentences through knowledge nodes in the financial knowledge graph, wherein the knowledge nodes have knowledge nodes of province A, city B, year C, tax data and GDP growth rate, and thus obtaining a plurality of key words.
An ambiguity analysis unit 30, configured to analyze whether an ambiguous keyword exists in the plurality of keywords, where the ambiguous keyword is a keyword that can find a plurality of knowledge nodes in a pre-established financial knowledge graph;
the financial knowledge map is a knowledge map which is created in advance and stored in a server, and the server judges whether ambiguous keywords of one keyword corresponding to a plurality of knowledge nodes exist in the financial knowledge map or not after acquiring the keywords. For example: the keyword X is a common public keyword, and knowledge nodes of a common public budget, a common public budget expenditure and a common public budget income are stored in the financial knowledge map, so that the keyword X in the financial problem statement has three knowledge nodes in the financial knowledge map, and the keyword X is an ambiguous keyword.
An ambiguity determining unit 40, configured to feed back, if yes, the knowledge nodes corresponding to the ambiguous keyword to the computer device, so that the user may select one of the knowledge nodes as a knowledge node corresponding to the ambiguous keyword;
if the server judges that one keyword corresponds to a plurality of knowledge nodes, the server calls the knowledge nodes in the form of options and sends the knowledge nodes to the computer equipment in the form of texts, so that the knowledge nodes can be conveniently presented on a screen of the computer equipment, and the effect of eliminating keyword ambiguity can be achieved by selecting one knowledge node by a user. For example: the server calls a general public budget, a general public budget expenditure and a general public budget income corresponding to the keyword X from the financial knowledge map, and feeds the general public budget, the general public budget expenditure and the general public budget income back to the computer equipment of the user to select any item, so that the effect of eliminating the ambiguity of the keyword X is achieved.
An invalid analysis unit 50 for analyzing whether there is an invalid keyword of a corresponding knowledge node that does not exist in the financial knowledge map among the plurality of keywords;
the server analyzes whether the keywords have invalid keywords which do not exist in the financial knowledge graph, and the mode is as follows: and searching whether knowledge nodes with the same meaning as the keywords exist in the financial knowledge graph. For example: the keyword "guangzhou city, tax payment data" appears in the financial problem statement, but only "guangzhou city" exists in the financial knowledge map but not "tax payment data", so the keyword "tax payment data" is an invalid keyword.
An invalid determining unit 60, configured to create a new knowledge node corresponding to the invalid keyword in the financial knowledge graph if the invalid keyword exists, perform corresponding internet search according to the invalid keyword, and use the searched keyword content as the node content of the new knowledge node;
if the server recognizes that the invalid keywords exist in the keywords, the invalid keywords are added into the financial knowledge graph as new knowledge nodes, but the invalid keywords are not associated with any content, so that the server searches for the content corresponding to the invalid keywords through the Internet to obtain node content, and the node content is used as the content of the invalid keywords, so that the training financial knowledge graph is expanded.
For example: and taking the keyword 'tax payment data' as an invalid keyword, adding a new knowledge node of the 'tax payment data' in the financial knowledge map, searching the content corresponding to the 'tax payment data' in the Internet, taking the searched content as the node content, and associating the node content with the knowledge node of the 'tax payment data'.
A path determining unit 70, configured to determine, in the financial knowledge graph, a knowledge node corresponding to a normal keyword in the plurality of keywords; confirming corresponding node search paths according to the front-back sequence of the ambiguous keywords, the invalid keywords and the normal keywords in the financial problem sentences and the knowledge nodes respectively corresponding to the ambiguous keywords, the invalid keywords and the normal keywords; the normal keywords are keywords which can find the unique knowledge node in a pre-established financial knowledge graph;
as shown in the above example, the sequence of the keywords in the financial problem statement is "A province, B city, C year, and tax data", so the server finds the node search path of the knowledge node corresponding to the "A province, B city, C year, and tax data" in the financial knowledge graph.
And the answer feedback unit 80 is configured to find a corresponding answer knowledge node in the financial knowledge graph according to the node search path, call answer content of the answer knowledge node, and feed back the answer content to the computer device.
As shown in the above example, the node search path of the keywords in the fiscal question sentence is "tax data" in a, B, C, so that the answer knowledge node is "tax data", and the server retrieves the answer content of "tax data" and feeds the answer content back to the computer device, wherein the answer content includes one or more of text answers, table answers, or picture answers.
Acquiring a financial problem statement input by a user operating computer equipment; semantically identifying a financial problem statement, and splitting the financial problem statement into a plurality of keywords; analyzing whether ambiguous keywords of a keyword corresponding to a plurality of knowledge nodes exist in a pre-established financial knowledge graph or not in the plurality of keywords; if yes, feeding the knowledge nodes back to the computer equipment so that the user can select one knowledge node as a knowledge node corresponding to the ambiguous keyword; analyzing whether the plurality of keywords have invalid keywords which do not exist in the financial knowledge graph; if so, creating a new knowledge node corresponding to the invalid keyword in the financial knowledge graph, performing corresponding internet search according to the invalid keyword, and taking the searched keyword content as the node content of the new knowledge node; confirming node search paths of knowledge nodes in a financial knowledge graph according to the front and back sequence of a plurality of keywords in a financial problem statement; finding a corresponding answer knowledge node in the financial knowledge map according to the node search path, calling the answer content of the answer knowledge node, and feeding the answer content back to the computer equipment; therefore, the problem that the existing machine has no function of eliminating ambiguity and the technical problem that the financial knowledge map is not formed and has low answer speed is solved.
Referring to fig. 4, a schematic block diagram of the interactive search answering apparatus according to an embodiment of the present invention is shown, wherein the financial knowledge map creating process is performed by:
the data acquisition unit 01 is used for acquiring text data input by a manager through a management terminal, wherein the text data comprises entity data and event data, the entity data is all knowledge nodes, and the event data is knowledge nodes with different association;
the manager makes a financial knowledge map through the server and manages users of the server; the entity class data comprises a first-order entity, a second-order entity, a third-order entity and the like, the event class data is a connection bridge of the entity class data, the first-order entity is connected with the second-order entity through an event, and the first-order entity is connected with the other first-order entity through an event.
For example: the first-order entity is 'A province', the second-order entity is 'B city', the event class data comprises an event type which is 'lower city', and therefore the events of 'A province' passing through 'lower city' are related to 'B city' in the financial knowledge map. The following description is required: the entity class data are knowledge nodes, and the event class data are conditions for associating the knowledge nodes.
A preliminary forming unit 02, configured to perform association connection on the knowledge nodes through event-like data to obtain a mesh-like preliminary knowledge graph;
the content searching unit 03 is configured to perform internet search on all knowledge nodes in the mesh-shaped preliminary knowledge graph to obtain knowledge node contents corresponding to all the knowledge nodes one to one;
and the content association unit 04 is used for associating all the knowledge nodes with the corresponding knowledge node content to obtain the financial knowledge graph.
After the server makes the preliminary knowledge graph, all knowledge nodes need to be filled with contents, and the method is as follows: searching all knowledge nodes one by one through the Internet to obtain the one-to-one corresponding knowledge node contents; and finally, performing association binding on the knowledge node content and the knowledge node corresponding to the knowledge node content to obtain the financial knowledge map.
In another embodiment, the financial knowledge-graph is regenerated within a predetermined time period in order to update the data, such as once a month.
In one embodiment, the dialogue-based search answering device further includes:
the same semantic identification unit is used for identifying whether the same knowledge nodes with the same semantic exist in all the knowledge nodes;
and the integration unit is used for integrating two or more same knowledge nodes into one knowledge node if the knowledge nodes are the same.
In the process of making the financial knowledge graph, if the server identifies the same knowledge nodes with the same semantics in all knowledge nodes, for example: and obtaining two knowledge nodes of the government of province A and the civil government of province A, identifying the semantics of the two knowledge nodes by the server, and integrating the two knowledge nodes into one knowledge node if the semantics are the same, namely the government of province A.
In one embodiment, the dialogue-based search answering device further includes:
an abbreviated keyword judgment unit for performing an internet search on the invalid keyword to judge whether the invalid keyword is an abbreviated keyword, if the invalid keyword exists;
the complete keyword acquisition unit is used for acquiring a complete keyword corresponding to the abbreviated keyword if the abbreviation keyword is found;
the searching unit is used for searching whether the knowledge node corresponding to the complete keyword exists in the financial knowledge map;
and the creating instruction generating unit is used for taking the complete key words as invalid key words if the complete key words do not exist, generating an instruction for creating new knowledge nodes corresponding to the invalid key words in the financial knowledge graph, and performing corresponding internet search according to the invalid key words.
For example: the financial problem statement input by the user is 'high and new enterprises in Shenzhen city', wherein the 'high and new enterprises' do not have corresponding knowledge nodes in the financial knowledge map, so that the 'high and new enterprises' are subjected to internet search, whether the 'high and new enterprises' are abbreviated keywords is judged, if the server searches for corresponding complete keywords through internet search, the complete keywords are obtained, and if the 'high and new enterprises' are subjected to internet search, the 'high and new technical enterprises' are obtained. And the server searches whether the knowledge node corresponding to the complete keyword exists in the financial knowledge graph or not, if not, the complete keyword is used as an invalid keyword, a new knowledge node for creating the corresponding invalid keyword in the financial knowledge graph is generated, and a creation instruction for searching the corresponding Internet is carried out according to the invalid keyword.
In one embodiment, the keyword recognition unit 20 includes:
the system comprises an unidentifiable confirming module, a keyword judging module and a keyword judging module, wherein the unidentifiable confirming module is used for splitting a financial problem statement into a plurality of words and identifying keywords corresponding to the words;
the professional term determining module is used for searching whether a professional term corresponding to the unrecognized word exists in a professional term library if the professional term cannot be recognized, wherein the professional term library is a database in which all the professional term is prestored;
and the term calling module is used for calling the professional term words as the keywords corresponding to the unrecognized words and taking the keywords corresponding to the words as a plurality of keywords obtained by splitting the financial problem sentence.
The above-mentioned keywords that cannot be identified are keywords that cannot be obtained by the server through semantic processing, so it is necessary to search the corresponding professional term in the professional term library.
Referring to fig. 5, a computer device, which may be a server and whose internal structure may be as shown in fig. 5, is also provided in the embodiment of the present application. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the computer designed processor is used to provide computational and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing data such as financial knowledge maps and the like. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a conversational search answer method.
The processor executes a dialogue-type search answering method, which comprises the following steps:
acquiring a financial problem statement input by a user through computer equipment;
semantically identifying a financial problem statement, and splitting the financial problem statement into a plurality of keywords;
analyzing whether ambiguous keywords exist in the plurality of keywords, wherein the ambiguous keywords are keywords of which a plurality of knowledge nodes can be found in a pre-established financial knowledge map;
if yes, feeding back a plurality of knowledge nodes corresponding to the ambiguous key words to the computer equipment so that the user can select one knowledge node as the knowledge node corresponding to the ambiguous key words;
analyzing whether the plurality of keywords have invalid keywords which do not have corresponding knowledge nodes in the financial knowledge graph;
if so, creating a new knowledge node corresponding to the invalid keyword in the financial knowledge graph, performing corresponding internet search according to the invalid keyword, and taking the searched keyword content as the node content of the new knowledge node;
determining knowledge nodes corresponding to normal keywords in a plurality of keywords in a financial knowledge graph; confirming corresponding node search paths according to the front-back sequence of the ambiguous keywords, the invalid keywords and the normal keywords in the financial problem sentences and the knowledge nodes respectively corresponding to the ambiguous keywords, the invalid keywords and the normal keywords; the normal keywords are keywords which can find the unique knowledge node in a pre-established financial knowledge graph;
and finding the corresponding answer knowledge node in the financial knowledge map according to the node search path, calling the answer content of the answer knowledge node, and feeding the answer content back to the computer equipment.
In one embodiment, the processor performs a financial knowledge map building method including:
acquiring text data input by a manager through a management terminal, wherein the text data comprises entity data and event data, the entity data is all knowledge nodes, and the event data is associated with different knowledge nodes;
performing association connection on the knowledge nodes through event data to obtain a meshed primary knowledge graph;
searching the internet for all knowledge nodes in the primary knowledge graph to obtain knowledge node contents corresponding to all the knowledge nodes one by one;
and associating all knowledge nodes with the corresponding knowledge node content to obtain the financial knowledge graph.
In one embodiment, the processor after executing the step of acquiring the text data input by the manager through the management terminal includes:
identifying whether the same knowledge nodes with the same semantics exist in all the knowledge nodes;
if yes, integrating two or more same knowledge nodes into one knowledge node.
In one embodiment, after the step of analyzing whether the plurality of keywords have invalid keywords corresponding to no knowledge nodes in the financial knowledge graph, the processor comprises:
if yes, carrying out Internet search on the invalid keywords to judge whether the invalid keywords are abbreviated keywords;
if yes, acquiring a complete keyword corresponding to the abbreviated keyword;
searching whether a knowledge node corresponding to the complete keyword exists in the financial knowledge graph;
and if the key words do not exist, taking the complete key words as invalid key words, generating an instruction for creating new knowledge nodes corresponding to the invalid key words in the financial knowledge graph, and performing corresponding internet search according to the invalid key words.
In one embodiment, the processor performs semantic recognition on the financial question statement, and the step of splitting the financial question statement into a plurality of keywords comprises:
splitting the financial problem statement into a plurality of words, and identifying keywords corresponding to the words respectively;
if the word is not identified, searching whether a professional term corresponding to the word which is not identified exists in a professional term library, wherein the professional term library is a database pre-storing all professional terms;
and calling the professional term words as keywords corresponding to the unrecognized terms, and taking the keywords corresponding to the plurality of terms as a plurality of keywords obtained by splitting the financial problem statement.
The above-mentioned keywords that cannot be identified are keywords that cannot be obtained by the server through semantic processing, so it is necessary to search the corresponding professional term in the professional term library.
Those skilled in the art will appreciate that the architecture shown in fig. 5 is only a block diagram of some of the structures associated with the present solution and is not intended to limit the scope of the present solution as applied to computer devices.
An embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of a conversational searching and answering method, and the method specifically includes:
the processor executes a dialogue-type search answering method, which comprises the following steps:
acquiring a financial problem statement input by a user through computer equipment;
semantically identifying a financial problem statement, and splitting the financial problem statement into a plurality of keywords;
analyzing whether ambiguous keywords exist in a plurality of keywords, wherein the ambiguous keywords are keywords of which a plurality of knowledge nodes can be found in a pre-established financial knowledge map;
if yes, feeding back a plurality of knowledge nodes corresponding to the ambiguous key words to the computer equipment so that the user can select one knowledge node as the knowledge node corresponding to the ambiguous key words;
analyzing whether the plurality of keywords have invalid keywords which do not have corresponding knowledge nodes in the financial knowledge graph;
if so, creating a new knowledge node corresponding to the invalid keyword in the financial knowledge graph, performing corresponding internet search according to the invalid keyword, and taking the searched keyword content as the node content of the new knowledge node;
determining knowledge nodes corresponding to normal keywords in a plurality of keywords in a financial knowledge graph; confirming corresponding node search paths according to the front-back sequence of the ambiguous keywords, the invalid keywords and the normal keywords in the financial problem sentences and the knowledge nodes respectively corresponding to the ambiguous keywords, the invalid keywords and the normal keywords; the normal keywords are keywords which can find the unique knowledge node in a pre-established financial knowledge graph;
and finding the corresponding answer knowledge node in the financial knowledge map according to the node search path, calling the answer content of the answer knowledge node, and feeding the answer content back to the computer equipment.
In one embodiment, the method for establishing the financial knowledge map performed by the processor comprises:
acquiring text data input by a manager through a management terminal, wherein the text data comprises entity data and event data, the entity data is all knowledge nodes, and the event data is associated with different knowledge nodes;
performing association connection on the knowledge nodes through event data to obtain a meshed primary knowledge graph;
searching the internet for all knowledge nodes in the primary knowledge graph to obtain knowledge node contents corresponding to all the knowledge nodes one by one;
and associating all knowledge nodes with the corresponding knowledge node content to obtain the financial knowledge graph.
In one embodiment, after the step of acquiring the text data input by the manager through the management terminal, the processor comprises:
identifying whether the same knowledge nodes with the same semantics exist in all the knowledge nodes;
if yes, integrating two or more same knowledge nodes into one knowledge node.
In one embodiment, after the step of analyzing whether the plurality of keywords have invalid keywords corresponding to no knowledge nodes in the financial knowledge graph, the processor comprises:
if yes, carrying out Internet search on the invalid keywords to judge whether the invalid keywords are abbreviated keywords;
if yes, acquiring a complete keyword corresponding to the abbreviated keyword;
searching whether a knowledge node corresponding to the complete keyword exists in the financial knowledge graph;
and if the keyword does not exist, taking the complete keyword as an invalid keyword, generating an instruction for creating a new knowledge node corresponding to the invalid keyword in the financial knowledge graph, and performing corresponding Internet search according to the invalid keyword.
In one embodiment, the processor performs semantic recognition on the financial question statement, and the step of splitting the financial question statement into a plurality of keywords includes:
splitting the financial problem statement into a plurality of words, and identifying keywords corresponding to the words respectively;
if the recognition is not available, searching whether a professional term corresponding to the unrecognized term exists in a professional term library, wherein the professional term library is a database pre-stored with all the professional terms;
and calling the professional term words as keywords corresponding to the unrecognized terms, and taking the keywords corresponding to the plurality of terms as a plurality of keywords obtained by splitting the financial problem statement.
The above-mentioned keywords that cannot be identified are keywords that cannot be obtained by the server through semantic processing, so it is necessary to search the corresponding professional term in the professional term library.
In conclusion, the server acquires the financial question sentences input by the user through the computer equipment; semantically identifying a financial problem statement, and splitting the financial problem statement into a plurality of keywords; analyzing whether ambiguous keywords exist in the plurality of keywords, wherein the ambiguous keywords are keywords of which a plurality of knowledge nodes can be found in a pre-established financial knowledge map; if yes, feeding back a plurality of knowledge nodes corresponding to the ambiguous keywords to the computer equipment so that the user can select one knowledge node as a knowledge node corresponding to the ambiguous keywords; analyzing whether the plurality of keywords have invalid keywords which do not have corresponding knowledge nodes in the financial knowledge graph; if so, creating a new knowledge node corresponding to the invalid keyword in the financial knowledge graph, performing corresponding internet search according to the invalid keyword, and taking the searched keyword content as the node content of the new knowledge node; determining knowledge nodes corresponding to normal keywords in a plurality of keywords in a financial knowledge graph; confirming a corresponding node search path according to the front-back sequence of the plurality of keywords in the financial problem sentence and the knowledge nodes corresponding to the keywords respectively; and finding the corresponding answer knowledge node in the financial knowledge graph according to the node search path, calling the answer content of the answer knowledge node, and feeding the answer content back to the computer equipment, so that the problem that the machine has no function of eliminating ambiguity and the technical problem that the financial knowledge graph is not formed and has low answer speed is solved.
It will be understood by those skilled in the art that all or part of the processes of the methods of the above embodiments may be implemented by hardware associated with instructions of a computer program, which may be stored on a non-volatile computer-readable storage medium, and when executed, may include processes of the above embodiments of the methods. Any reference to memory, storage, database or other medium provided herein and used in the examples may include non-volatile and/or volatile memory. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), double-rate SDRAM (SSRSDRAM), Enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, apparatus, article, or method. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, apparatus, article, or method that includes the element.
The above description is only for the preferred embodiment of the present application and not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings of the present application, or which are directly or indirectly applied to other related technical fields, are intended to be included within the scope of the present application.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (10)
1. A conversational search response method, comprising:
acquiring a financial problem statement input by a user through computer equipment;
semantically identifying the financial problem statement, and splitting the financial problem statement into a plurality of keywords;
analyzing whether ambiguous keywords exist in the plurality of keywords, wherein the ambiguous keywords are keywords which can be found out from a plurality of knowledge nodes in a pre-established financial knowledge graph;
if yes, feeding back a plurality of knowledge nodes corresponding to the ambiguous keyword to the computer equipment so that the user can select one knowledge node as the knowledge node corresponding to the ambiguous keyword;
analyzing whether the plurality of keywords have invalid keywords which do not have corresponding knowledge nodes in the financial knowledge graph;
if yes, a new knowledge node corresponding to the invalid keyword is created in the financial knowledge graph, corresponding internet search is carried out according to the invalid keyword, and the searched keyword content is used as the node content of the new knowledge node;
determining knowledge nodes corresponding to normal keywords in the plurality of keywords in the financial knowledge graph; confirming corresponding node search paths according to the front-back sequence of the ambiguous keywords, the invalid keywords and the normal keywords in the financial problem sentence and the knowledge nodes respectively corresponding to the ambiguous keywords, the invalid keywords and the normal keywords; the normal keywords are keywords which can find the only knowledge node in the pre-established financial knowledge graph;
and finding a corresponding answer knowledge node in the financial knowledge map according to the node search path, calling the answer content of the answer knowledge node, and feeding the answer content back to the computer equipment.
2. The conversational search response method of claim 1, wherein the financial knowledge base establishing method comprises:
acquiring text data input by a manager through a management terminal, wherein the text data comprises entity data and event data, the entity data is all knowledge nodes, and the event data is associated with different knowledge nodes;
performing association connection on the knowledge nodes through the event data to obtain a meshed primary knowledge graph;
searching the internet for all knowledge nodes in the primary knowledge graph to obtain knowledge node contents corresponding to all the knowledge nodes one by one;
and associating all knowledge nodes with the corresponding knowledge node content to obtain the financial knowledge graph.
3. The interactive search answering method according to claim 2, wherein said step of obtaining the text data inputted by the manager through the management terminal is followed by:
identifying whether the same knowledge nodes with the same semantics exist in all the knowledge nodes;
and if so, integrating two or more same knowledge nodes into one knowledge node.
4. The interactive search answering method according to claim 1, wherein said step of analyzing said plurality of keywords for invalid keywords in said financial knowledge graph without corresponding knowledge nodes comprises:
if yes, carrying out Internet search on the invalid keyword to judge whether the invalid keyword is an abbreviated keyword;
if yes, acquiring a complete keyword corresponding to the abbreviated keyword;
searching whether a knowledge node corresponding to the complete keyword exists in the financial knowledge graph;
and if the complete key words do not exist, taking the complete key words as the invalid key words, generating instructions for creating new knowledge nodes corresponding to the invalid key words in the financial knowledge graph, and performing corresponding internet search according to the invalid key words.
5. The interactive search answering method according to claim 1, wherein the semantic recognition of the financial question sentence, and the step of splitting the financial question sentence into a plurality of keywords comprises:
splitting the financial problem statement into a plurality of words, and identifying keywords corresponding to the words respectively;
if the recognition is not available, searching whether a professional term corresponding to the unrecognized term exists in a professional term library, wherein the professional term library is a database in which all the professional terms are prestored;
and calling the professional term as a keyword corresponding to the unrecognized term, and taking the keyword corresponding to each of the plurality of terms as a plurality of keywords obtained by splitting the financial problem statement.
6. A conversational search answering device, comprising:
the statement acquisition unit is used for acquiring financial problem statements input by a user through computer equipment;
the keyword identification unit is used for semantically identifying the financial problem statement and splitting the financial problem statement into a plurality of keywords;
the ambiguity analysis unit is used for analyzing whether an ambiguity keyword exists in the plurality of keywords, wherein the ambiguity keyword is a keyword which can search a plurality of knowledge nodes in a pre-established financial knowledge graph;
an ambiguity determining unit, configured to feed back, if yes, the plurality of knowledge nodes corresponding to the ambiguous keyword to the computer device, so that the user can select one of the knowledge nodes as a knowledge node corresponding to the ambiguous keyword;
an invalid analysis unit, configured to analyze whether an invalid keyword in the plurality of keywords exists in the financial knowledge graph, where the invalid keyword does not exist in a corresponding knowledge node;
an invalid determining unit, configured to create a new knowledge node corresponding to the invalid keyword in the financial knowledge graph if the invalid keyword exists, perform corresponding internet search according to the invalid keyword, and use the searched keyword content as the node content of the new knowledge node;
a path determining unit, configured to determine, in the financial knowledge map, a knowledge node corresponding to a normal keyword in the plurality of keywords; confirming corresponding node search paths according to the front-back sequence of the ambiguous keywords, the invalid keywords and the normal keywords in the financial problem sentence and the knowledge nodes respectively corresponding to the ambiguous keywords, the invalid keywords and the normal keywords; the normal keywords are keywords of which the only knowledge node can be found in the pre-established financial knowledge graph;
and the answer feedback unit is used for finding the corresponding answer knowledge node in the financial knowledge map according to the node search path, calling the answer content of the answer knowledge node and feeding the answer content back to the computer equipment.
7. The interactive search response device according to claim 6, further comprising:
the system comprises a data acquisition unit, a data processing unit and a data processing unit, wherein the data acquisition unit is used for acquiring text data input by a manager through a management terminal, the text data comprises entity data and event data, the entity data are all knowledge nodes, and the event data are associated with different knowledge nodes;
the preliminary forming unit is used for performing association connection on the knowledge nodes through the event data to obtain a meshed preliminary knowledge graph;
the content searching unit is used for searching the internet for all knowledge nodes in the primary knowledge graph in the shape of a mesh to obtain knowledge node contents corresponding to all the knowledge nodes one by one;
and the content association unit is used for associating all the knowledge nodes with the corresponding knowledge node content to obtain the financial knowledge graph.
8. The interactive search response device according to claim 7, further comprising:
the same semantic identification unit is used for identifying whether the same knowledge nodes with the same semantic exist in all the knowledge nodes;
and the integration unit is used for integrating two or more same knowledge nodes into one knowledge node if the knowledge nodes are the same.
9. A computer device comprising a memory and a processor, the memory having stored therein a computer program, characterized in that the processor, when executing the computer program, carries out the steps of the conversational search response method of any of claims 1 to 5.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the dialogue-based search response method according to any one of claims 1 to 5.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910464157.2A CN110347810B (en) | 2019-05-30 | 2019-05-30 | Dialogue type search answering method, device, computer equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910464157.2A CN110347810B (en) | 2019-05-30 | 2019-05-30 | Dialogue type search answering method, device, computer equipment and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110347810A CN110347810A (en) | 2019-10-18 |
CN110347810B true CN110347810B (en) | 2022-08-19 |
Family
ID=68174506
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910464157.2A Active CN110347810B (en) | 2019-05-30 | 2019-05-30 | Dialogue type search answering method, device, computer equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110347810B (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111209389B (en) * | 2019-12-31 | 2023-08-11 | 天津外国语大学 | Movie story generation method |
CN112818092B (en) * | 2020-04-20 | 2023-08-11 | 腾讯科技(深圳)有限公司 | Knowledge graph query statement generation method, device, equipment and storage medium |
CN112115276B (en) * | 2020-09-18 | 2024-05-24 | 平安科技(深圳)有限公司 | Intelligent customer service method, device, equipment and storage medium based on knowledge graph |
CN112396898A (en) * | 2020-11-09 | 2021-02-23 | 江苏创设未来教育发展有限公司 | Skill learning practical training education system based on artificial intelligence |
CN117633540B (en) * | 2024-01-25 | 2024-04-30 | 杭州阿里云飞天信息技术有限公司 | Sample data construction method and device |
Family Cites Families (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105431839A (en) * | 2013-03-15 | 2016-03-23 | 罗伯特·哈多克 | Intelligent internet system with adaptive user interface providing one-step access to knowledge |
CN103914548B (en) * | 2014-04-10 | 2018-01-09 | 北京百度网讯科技有限公司 | Information search method and device |
US10606846B2 (en) * | 2015-10-16 | 2020-03-31 | Baidu Usa Llc | Systems and methods for human inspired simple question answering (HISQA) |
CN105843875B (en) * | 2016-03-18 | 2019-09-13 | 北京光年无限科技有限公司 | A kind of question and answer data processing method and device towards intelligent robot |
US10169470B2 (en) * | 2016-04-11 | 2019-01-01 | Rovi Guides, Inc. | Systems and methods for identifying a meaning of an ambiguous term in a natural language query |
CN108170662A (en) * | 2016-12-07 | 2018-06-15 | 富士通株式会社 | The disambiguation method of breviaty word and disambiguation equipment |
CN106919655B (en) * | 2017-01-24 | 2020-05-19 | 网易(杭州)网络有限公司 | Answer providing method and device |
CN108153901B (en) * | 2018-01-16 | 2022-04-19 | 北京百度网讯科技有限公司 | Knowledge graph-based information pushing method and device |
CN109145102B (en) * | 2018-09-06 | 2021-02-09 | 杭州安恒信息技术股份有限公司 | Intelligent question answering method and knowledge graph system construction method, device and equipment thereof |
CN109492111B (en) * | 2018-09-19 | 2023-05-30 | 平安科技(深圳)有限公司 | Shortest path query method, shortest path query system, computer device and storage medium |
CN109614476A (en) * | 2018-12-11 | 2019-04-12 | 平安科技(深圳)有限公司 | Customer service system answering method, device, computer equipment and storage medium |
CN109753557B (en) * | 2018-12-26 | 2021-01-01 | 出门问问信息科技有限公司 | Answer output method, device, equipment and storage medium of question-answering system |
-
2019
- 2019-05-30 CN CN201910464157.2A patent/CN110347810B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN110347810A (en) | 2019-10-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110347810B (en) | Dialogue type search answering method, device, computer equipment and storage medium | |
US10755046B1 (en) | Applied artificial intelligence technology for conversational inferencing | |
US9582757B1 (en) | Scalable curation system | |
US8170866B2 (en) | System and method for increasing accuracy of searches based on communication network | |
US9213758B2 (en) | Method and apparatus for responding to an inquiry | |
CN110675288A (en) | Intelligent auxiliary trial method and device, computer equipment and storage medium | |
CN112052324A (en) | Intelligent question answering method and device and computer equipment | |
CN109840255B (en) | Reply text generation method, device, equipment and storage medium | |
CN108538294B (en) | Voice interaction method and device | |
NZ541471A (en) | System for predicting speech recognition accuracy and development for a dialog system | |
CN110019304B (en) | Method for expanding question-answering knowledge base, storage medium and terminal | |
CN110704571A (en) | Court trial auxiliary processing method, trial auxiliary processing device, equipment and medium | |
CN110781280A (en) | Knowledge graph-based voice assisting method and device | |
CN111223476B (en) | Method and device for extracting voice feature vector, computer equipment and storage medium | |
CN110765342A (en) | Information query method and device, storage medium and intelligent terminal | |
CN111782800A (en) | Intelligent conference analysis method for event tracing | |
CN110019305B (en) | Knowledge base expansion method, storage medium and terminal | |
WO2020233381A1 (en) | Speech recognition-based service request method and apparatus, and computer device | |
CN113343108A (en) | Recommendation information processing method, device, equipment and storage medium | |
CN111368029A (en) | Interaction method, device and equipment based on intention triples and storage medium | |
CN110609618A (en) | Man-machine conversation method and device, computer equipment and storage medium | |
CN111813916B (en) | Intelligent question-answering method, device, computer equipment and medium | |
CN110931002B (en) | Man-machine interaction method, device, computer equipment and storage medium | |
US20060136195A1 (en) | Text grouping for disambiguation in a speech application | |
CN111506701B (en) | Intelligent query method and related device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20230704 Address after: No. 888, Huanhu West 2nd Road, Lingang New District, China (Shanghai) pilot Free Trade Zone, Pudong New Area, Shanghai, 200120 Patentee after: Shanghai Ruyi Information Technology Co.,Ltd. Address before: 400010 38 / F, 39 / F, unit 1, 99 Wuyi Road, Yuzhong District, Chongqing Patentee before: CHONGQING FINANCIAL ASSETS EXCHANGE Co.,Ltd. |